Cloud deployment platform Railway disclosed this week that an unauthorized party may have broken into systems storing customer environment variables, triggering a platform-wide pause on new deployments and an urgent push for developers to rotate every API key, database credential, and registry login stored on the service. The company later estimated that at most 1.85 percent of all workspaces — roughly 100,000 — may have had environment variables compromised, though it said the intruder never accessed Railway user accounts themselves.
What Railway Says Happened
Railway later said it became aware of a sustained campaign at 19:53 UTC on October 20, before disclosing anomalous activity on Wednesday, according to a security disclosure posted to its blog. The company said an unauthorized third party may have accessed encrypted environment variables and, for some of them, obtained the decrypted contents — variables that can include secrets such as API keys, database credentials, or registry logins. At that point, Railway said its initial analysis suggested the impact might be limited to a subset of users, and it strongly urged customers to immediately roll all stored environment variables and any private registry credentials.
Railway paused build infrastructure to prioritize user safety and mitigate attack vectors while it investigated, telling users that until builds were restored, they would not be able to roll out new versions. As first reported by OfficeChai, the freeze briefly left developers unable to even deploy the rotated credentials Railway had just told them to generate.
Scope Narrows, But Not Fully
By the next day, Railway's preliminary report offered a clearer — though still partial — picture. The company said at most 1.85 percent of all workspaces, or approximately 100,000 workspaces, had environment variables potentially compromised, and it stated plainly that the attacker did not access Railway user accounts as part of the incident. The same report noted that all existing user deployments had remained online and running throughout, and that builds for all tiers had been restored.
Railway said it proactively worked with GitHub, GitLab, Anthropic, OpenAI, OpenRouter, SpaceXSI, Stripe, and other vendors to revoke potentially compromised keys before they could be misused. By the time Railway posted an incident status update, it declared the incident resolved, with builds, deployments, the API, and the dashboard all operating normally. Still, the same status update cautioned that workspaces outside the higher-risk groups Railway had identified were not confirmed safe from compromise — a line that leaves plenty of developers outside the initial spotlight uncertain about their own exposure.
A Hunt for AI API Keys
What makes the Railway incident notable isn't just its scale but its apparent target. The company's preliminary report said forensic evidence indicated the attack was attempting to obtain AI-provider API keys to perform distillation attacks — a technique where stolen access to a premium AI model is used to cheaply train or extract capabilities from a competing model. As reported by TechCrunch, Railway characterized the breach as likely part of a broader attack targeting platforms with the aim of compromising AI inference providers.
That framing lines up with other recent incidents. Days earlier, TechCrunch reported that hackers hit the AI coding tool OpenCode, running up charges on users' AI accounts in a breach the outlet described as landing amid a run of AI-enabled cyber incidents. Separately, the OpenCode Console incident affected 1.4 percent of Console users, all of whom had an outside AI provider linked to their accounts. The Hacker News also covered the incident. TechCrunch also noted that OpenCode said hackers took more than 1 percent of its customers' API keys and other secrets in a similar attack.
Not an Isolated Pattern
Railway and OpenCode aren't the only platforms grappling with this kind of exposure this month. AI infrastructure company Fireworks disclosed in a security bulletin that an unauthorized third party accessed environment variables used in evaluation jobs in a limited number of customer accounts, and that it believed API keys and tokens stored as secrets were captured in certain affected accounts. Security firm GitGuardian, meanwhile, wrote on its blog that the Railway attack pattern echoed a Vercel incident from April, suggesting attackers have been probing cloud deployment platforms for stored secrets as a recurring strategy.
For now, Railway has not disclosed how the unauthorized access occurred, nor has it named which specific upstream providers had exposed keys. Its guidance to customers remains focused on fundamentals: rotate every environment variable, including API keys, and review service logs for unusual activity. Developers whose workspaces fall outside the groups Railway has flagged as higher-risk are left in an uncomfortable middle ground — not confirmed compromised, but not confirmed clean either.
